INDEX

Explanations

publishers and publication information

oai_token-act-pair · gpt-3.5-turbo

references to publishers and publication details

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.11.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

aucuses

-0.92

ndum

-0.81

auga

-0.80

 migr

-0.75

 drills

-0.75

philis

-0.74

xual

-0.73

hya

-0.72

merce

-0.70

 drilled

-0.69

POSITIVE LOGITS

 Warner

1.43

 Paramount

1.25

 Ubisoft

1.16

 Universal

1.12

MGM

1.11

 Sony

1.10

 Activision

1.09

 Penguin

1.09

 Pixar

1.08

 Valiant

1.07

Activations Density 0.208%

publishers and publication information

references to publishers and publication details

No Comments

No Known Activations